The first question to ask ourselves before starting a project is “who will use a product we develop?” The more we know about them, the better it is for developing and testing process.
First of all, we are to figure out the kind of target audience of the chatbot: who are the customers and the end users?
To find out more about target audience we use a statistical approximation of a client’s web-page traffic. We assume that the web-page audience will be the chatbot audience as well. Of course, there are specific stats for different messenger platforms on which chatbot will be developed. These stats also help us learn about target users.
The statistics give us useful information about the geography of users (country, language). Based on this information we can:
add to our chatbot some personality characteristics, some behavioral things according to users folk or traditions we’ve got to know;
create a multilanguage chatbot to cover more users (audience);
propose different messenger platforms for different target audiences;
select appropriate countries for stats for cross-platform and cross-device testing in order to get the best test-coverage.
Also, we get information about gender and age of users. It helps us a lot:
to create a chatbot personality and conversational UX (there are many types of research revealing how men and women chat and talk within messengers. For example, women prefer to use emoticons.);
to add an appropriate content to stay, and engage users growing bot retention according to age and gender;
to provide useful features that solve relevant user problems;
to create and describe suitable test cases according to behavioral characteristics of using messenger platform.
The answers to this first question simply and quickly give us useful information to create and provide high-quality and high-retention chatbots you may see in our portfolio.
If you weren’t thinking about chatbots for your organization before, you should be now. 110000 bots on Messenger can’t be wrong. Let’s build your bot!
Also published on Medium.comments powered by HyperComments